Liquid staking is safe in the ways people worry about and risky in the ways they don't. Five real risks of an LST, each with a check you can run yourself before depositing.

Short answer: The risk most people ask about — "can the protocol just take my coins?" — is the one that is easiest to rule out, and you can rule it out in about four minutes with a block explorer. The risks that actually cost liquid staking users money are duller: you cannot exit instantly, the token can trade below its backing while you wait, and the rewards depend on a contract that someone has to keep funded.
Here are all five, in the order they are likely to affect you, each with a check you can run yourself.
Every liquid staking token is a claim on something illiquid, wrapped in something liquid. Every risk below is a consequence of that one sentence.
You deposit a proof-of-stake asset — SUI, EGLD, XRD. You receive a liquid staking token (LST) representing that deposit. The underlying stays delegated to validators and earns rewards; your LST stays in your wallet and can be traded, lent or farmed.
On JewelSwap this runs as a dual-token model. Depositing SUI mints JWLSUI, the base token. Stake JWLSUI and you receive SJWLSUI, which appreciates against JWLSUI as rewards accrue. The same shape applies on MultiversX (JWLEGLD / SJWLEGLD) and Radix (JWLXRD / SJWLXRD).
The part that matters for risk: redemption is not instant. Converting back to the underlying runs through a 10-day unbonding period, mirroring the chain's own validator unstaking delay. During that window you hold a claim NFT rather than the asset.
That 10-day window is the source of risks one and two.
This is the risk that actually materialises, on every LST, on every chain.
If you want the underlying asset today, you have two options. Wait out the unbonding period and redeem at par. Or sell your LST into a liquidity pool right now and accept whatever the pool offers. When a lot of people want out at once, the second number drops — sometimes a long way below the first.
This is not a malfunction. It is the market pricing your impatience. stETH traded under ETH through 2022 while every single ETH behind it was fully accounted for.
We have published our own worked example rather than describing someone else's: JWLEGLD's full on-chain depeg history, with the transactions.
A discount on a liquid staking token is not evidence of insolvency. It is evidence that the exit door is narrower than the room.
How to check it yourself — before you deposit, not after:
The gap is your real cost of leaving in a hurry. If a 1-token quote looks fine but a 500-token quote is 40% worse, your position is only as liquid as the pool is deep.
⚠ One technical trap, because we fell into it ourselves. If the pool is a stableswap pool with a high amplification factor, the reserve ratio is not the price. We once read a pool's reserves and recorded our own peg as 3.6% when the router was quoting 82%. Always quote the router, never the reserves.
Every LST protocol says its token is backed. The word does more work in some places than others, so read the mechanism rather than the adjective.
Ours is a concrete example. JewelSwap's documentation states that JWLSUI is 1:1 backed, and also that the protocol can mint up to 1.1 JWLSUI per deposited SUI through Protocol-Owned Liquidity — tokens minted against the protocol's own balance sheet to seed pool depth, not against a user deposit.
That is a legitimate and common design. Deeper pools directly reduce risk 1. But it means "1:1 backed" describes the user's redemption claim, not the ratio of total supply to total reserves. Those are different numbers, and any protocol that blurs them is worth more scrutiny, including this one.
How to check it yourself:
If supply materially exceeds the underlying reserves, ask what the difference is and who is on the hook for it. A good protocol will answer with a mechanism. A bad one will repeat the adjective.
Ask "backed by what, held where, redeemable by whom, in how many days?" Four answers. If any one of them is missing, you have found the risk.
Your LST earns whatever the validators behind it earn. Underperforming or misconfigured validators mean lower rewards; on chains with slashing, validator misbehaviour can destroy a portion of the delegated stake.
This risk is mostly delegated away from you — and that is exactly why it is worth inspecting who it is delegated to. On JewelSwap, validator delegation is decided through Gauge governance, where stakers vote on distribution.
How to check it yourself: find out who chooses the validator set, and how concentrated it is. One operator holding most of the delegation is a single point of failure regardless of how good that operator is. Ask whether the protocol has a stated policy for removing an underperforming validator, and how fast that can actually happen.
Here is the risk everyone asks about first, and it is genuinely checkable.
Two questions decide most of it.
Who can upgrade the contracts? If a single externally-owned account can push a new implementation, the security of your deposit is the security of one private key. A multisig is materially better — but only if the signers are actually independent. A 3-of-5 where three keys sit on one laptop is a 1-of-1 wearing a costume.
What can admin functions actually do? Pausing deposits is normal and often protective. Moving user funds is not.
How to check it yourself:
That last point is subtler than it sounds. Capabilities that look locked down often are not — a wrapper function that re-exposes a privileged action makes the "soulbound" label meaningless. If you cannot establish who controls upgrades from public data in a few minutes, treat that as the answer.
An audit tells you what the code did on the day it was read. Key control tells you who can change it tomorrow. The second one matters more.
The full version of this check, applicable to any protocol rather than just an LST, is in how to check a DeFi protocol before you deposit.
This one gets almost no coverage and it is a real failure mode.
The staking contract and the contract that pays rewards are usually separate. The rewarder holds a balance and emits at some rate. If emissions outrun the balance, reward claims start failing — and depending on how the contract is written, the functions that fail may include the ones you need to withdraw.
The tell is arithmetic: compare the rewarder's current balance against its daily emission rate. A rewarder holding two days of emissions is not a rewarder, it is a countdown.
How to check it yourself: find the reward distributor's address, read its balance and its emission rate, and divide. Then check that claim transactions are actually succeeding — a run of reverts on a reward contract is visible in its recent transaction history, and it is visible before anyone announces it.
Liquid staking is a mature mechanism. The contract risk is real but bounded and inspectable. The risks that reliably cost people money are the structural ones: you cannot exit instantly, and the instant exit is priced.
A reasonable position:
The protocols that survive are not the ones that never trade at a discount. They are the ones that tell you why they did.
The contract and custody risks are bounded and independently checkable. The risks that actually cost people money are structural: a fixed unbonding period you cannot skip, and an exit price set by pool depth rather than by backing. Liquid staking is safe in the sense that the mechanism works as designed; it is risky in the sense that the design does not guarantee you can leave at par whenever you want.
Exit liquidity. If you need the underlying asset before the unbonding period ends, your only option is selling into a pool, and that price is set by how deep the pool is relative to your position. This is the risk that materialises most often, on every chain.
Yes, and it does not imply insolvency. An LST is not a stablecoin and is not designed to hold a hard 1:1 price. A discount prices the wait, not the backing — stETH traded below ETH through 2022 while every ETH behind it was accounted for.
On JewelSwap the unbonding period is 10 days across MultiversX, Sui and Radix, mirroring each chain's own validator unstaking delay. During that window you hold a transferable claim NFT rather than the underlying asset.
On chains with slashing, validator misbehaviour can destroy part of the delegated stake, and that loss passes through to holders. This is why validator selection and concentration matter: one operator holding most of the delegation is a single point of failure regardless of that operator's quality.
Check four things: the realistic exit quote at your position size, total supply against reserves held, who controls contract upgrades and whether the multisig signers are genuinely independent, and whether the rewards contract has more than a few days of runway against its emission rate. All four are answerable from public data.
This article is educational and is not financial advice. Verify any protocol's contract state yourself before depositing.